Approach to Video Stream Quality Requirements Justification for FPV Control of Unmanned Systems

Currently, hybrid orbital-ground communication networks play a key role in various sectors of the Russian economy, an important part of which are unmanned systems based on FPV control. One of the key elements of such systems is information exchange channels. The main criterion for the success of their objectives is to meet the requirements for the quality of the transmitted video stream, which makes it particularly relevant to substantiate these requirements. The purpose of this article is to present an approach to the quantitative justification of the requirements for the quality of the transmitted video stream. The essence of the presented approach lies in the fact that the required values of the transmitted image quality indicators, used to justify acceptable neural network codecs, are determined by analyzing various video streams, on the basis of which the impacts on the control of unmanned systems that allowed to achieve the goals of their functioning were formed. The main stages of the approach and their logical interrelation are considered. The proposed approach is based on the use of methods of statistical and risk analysis, the theory of experiment planning and probability theory.Scientific novelty of the proposed approach lies in the fact that the requirements to the quality of the video stream transmitted from the unmanned system to the operator, used to justify acceptable neural network codecs, are calculated by analyzing the totality of video streams that allowed to achieve the goals of its functioning, which allows to get rid of the subjectivism inherent in the expert methods currently used to solve this problem.Theoretical significance of the proposed approach lies in the fact that the necessity and possibility of not only substantiating the requirements for the quality indicators of the transmitted video stream for FPV control, but also their correction has been proved. Practical significance. The requirements for the quality indicators of the transmitted video stream obtained using the proposed approach can be further used to justify optimal design solutions for creating information exchange channels between unmanned systems and the operator, which will significantly increase the efficiency of using these systems.

hybrid netwoks, unmanned vehicles, FPV control, the quality of the restored video stream, risk, nuclear probability density estimation method, requirements, neural network codecs